Instructions to use DrIAmed/Voxtral-Darija-ASR with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use DrIAmed/Voxtral-Darija-ASR with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("mistralai/Voxtral-Mini-3B-2507")
model = PeftModel.from_pretrained(base_model, "DrIAmed/Voxtral-Darija-ASR")

Transformers

How to use DrIAmed/Voxtral-Darija-ASR with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="DrIAmed/Voxtral-Darija-ASR")

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("DrIAmed/Voxtral-Darija-ASR", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use DrIAmed/Voxtral-Darija-ASR with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "DrIAmed/Voxtral-Darija-ASR"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DrIAmed/Voxtral-Darija-ASR",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/DrIAmed/Voxtral-Darija-ASR

SGLang

How to use DrIAmed/Voxtral-Darija-ASR with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "DrIAmed/Voxtral-Darija-ASR" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DrIAmed/Voxtral-Darija-ASR",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "DrIAmed/Voxtral-Darija-ASR" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "DrIAmed/Voxtral-Darija-ASR",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use DrIAmed/Voxtral-Darija-ASR with Docker Model Runner:
```
docker model run hf.co/DrIAmed/Voxtral-Darija-ASR
```

Voxtral-Darija-ASR

This model is a fine-tuned version of mistralai/Voxtral-Mini-3B-2507 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 2.8675

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0002
train_batch_size: 4
eval_batch_size: 1
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 50
num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss
2.7092	0.7299	100	2.7637
2.1263	1.4599	200	2.5980
1.5760	2.1898	300	2.4979
1.6346	2.9197	400	2.4325
1.0920	3.6496	500	2.5351
0.6005	4.3796	600	2.6709
0.2347	5.1095	700	2.8547
0.2175	5.8394	800	2.8675

Framework versions

PEFT 0.18.1
Transformers 5.3.0
Pytorch 2.4.1+cu124
Datasets 2.18.0
Tokenizers 0.22.2

Downloads last month: 2

Model tree for DrIAmed/Voxtral-Darija-ASR

Base model

mistralai/Voxtral-Mini-3B-2507

Adapter

(12)

this model